Endocrine Abstracts (2019) 65 P285 | DOI: 10.1530/endoabs.65.P285

The argument for growth hormone day curve testing in acromegaly; significant discrepancy between mean growth hormone and random growth hormone levels

Ultan Healy, Shoaib Kahn, Alex Vincent, Jacinto Martinez, Christine J H May, John Wass, Brian Shine, Bahram Jafar-Mohammadi & Aparna Pal


Oxford Centre for Diabetes, Endocrinology and Metabolism, Oxford University Hospitals NHS Foundation Trust, Oxford, UK


Endocrine Society guidelines recommend random growth hormone (RGH) <1 mcg/l as indicative of biochemical control of acromegaly. Growth hormone (GH) control may also be determined using the mean GH (MGH) of a growth hormone day curve (GHDC). We report a retrospective analysis of 461 consecutive GHDCs, from 121 patients with treated acromegaly, performed in a single centre between 2009 and 2019. Each GHDC contained 7–9 GH measurements (mean 8.9) taken at regular intervals over 9-hours. Each individual GH measurement was treated as a RGH (n=4113) and compared against the MGH of its own curve. Contemporaneous IGF1 levels, expressed as a fraction of the sex- and age-adjusted upper limit of normal, were recorded. GH was measured on a Siemens Immulite 2000 assay. IGF-1 was measured on the Immulite 2000 assay prior to 2013, and subsequently on an IDS iSys assay. Sequential RGH levels displayed significant variance (mean±SD; 0.57±2.64). Variance was greater if MGH ≥1 mcg/l (1.12±3.88) compared to MGH <1 mcg/l (0.13±0.27, P=0.00005). Disagreement (defined as either RGH <1 mcg/l and MGH ≥1 mcg/l, or RGH ≥1 mcg/l and MGH <1 mcg/l) was observed in 191/461 (41.4%) GHDCs, involving 510/4113 (12.4%) RGHs. 271/4113 (6.6%) RGHs significantly underestimated MGH, with 75/4113 (1.8%) of RGHs <1 mcg/l despite corresponding MGH ≥1.5 mcg/l. RGH ≤0.3 mcg/l had a predictive value of 94.9% (444/468) for MGH <1 mcg/l. 204/4113 (5.0%) RGHs significantly overestimated MGH, with 70/4113 (1.7%) of RGHs ≥1.5 mcg/l despite corresponding MGH <1 mcg/l. RGH ≥2 mcg/l had a predictive value of 96.2% (842/875) for MGH ≥1 mcg/l. Rates of GH discordance (GH ≥1 mcg/l, IGF1 ≤ULN) and IGF1 discordance (IGF1 >ULN, GH <1 mcg/l) were similar for RGH (644/4113, 15.6% and 733/4113, 17.8%) and MGH (70/461, 15.2%, P=0.1367 and 82/461, 17.8%, P=0.986). We conclude that RGH displays significant variance, is frequently (12.4%) in disagreement with MGH, and often significantly overestimates (5.0%) or underestimates (6.6%) GH exposure.